KR100930594B1 - The system for capturing 2d facial image and extraction method of face feature points thereof - Google Patents

Abstract

PURPOSE: A facial image photographing device and a facial feature point detecting method thereof are provided to decide whether a triangle which connects central points of the pupils with a central point of the mouth satisfies a symmetric condition in a front facial image of a photographed target person. CONSTITUTION: A first detector(130) detects a facial area of a target person from a front facial image by using color information for a face. The first detector detects both pupils and mouth in the detected facial area. The first detector detects first/second spots which are central points of both pupils and a third spot which is a central point of the mouth. The detected first to third spots are connected together to form a triangle. A controller(150) decides whether one side of the triangle is parallel with the ground.

Description

Facial imaging apparatus and its facial feature point detection method {The system for capturing 2D facial image and extraction method of face feature points

The present invention relates to a facial imaging apparatus and a facial feature point detecting method thereof, and more particularly, to a facial image capturing apparatus and a facial feature point detection method thereof for photographing a front image of a subject.

In general, feature points in images of a specific person are used in various technical fields. In particular, the feature points extracted from the facial image of the target person may be usefully used in facial recognition technology to determine the identity of the target person through the image.

Feature points in these facial images can also be used in the medical field. With the development of current science and technology, technologies are being developed that can measure the degree of health or abnormality by analyzing images of patients at a distance without facing doctors and patients directly. In addition, it is possible to apply facial diagnosis in a home-care system or a ubiquitous healthcare system.

As described above, it is important to accurately extract the feature points in order to perform diagnosis through the face image of the patient. If the face is tilted when taking a picture, the accuracy of feature point extraction is degraded. In this case, there is a problem that it is difficult to properly diagnose through facial images. Accordingly, there is a need for a facial image capturing apparatus capable of correcting a posture of a target person and capturing a facial image of a correct posture.

In addition, it is possible to apply to the generation and animation of two-dimensional image-based three-dimensional personal face model using a front face image of a person can be used for games and entertainment using a three-dimensional face in places such as mobile phones.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an object of the present invention is to capture a front image of a target person and correct a posture, so that a face image photographing apparatus capable of obtaining a front image having a correct posture and facial feature points thereof It is for providing a detection method.

In addition, another object of the present invention is to detect the facial feature point in only one of the left region and the right region symmetrical with respect to the centerline of the front image, and to detect the face matching point matching the other region faster, more accurately A facial imaging apparatus capable of detecting facial feature points and a method for detecting facial feature points thereof are provided.

According to an aspect of the present invention, there is provided a facial image capturing apparatus comprising: a camera for photographing a front image of a target person in real time, a display unit for outputting a front image captured by the camera, and color information. The first and second points, which are the center points of both eyes, are detected in the front image, and the first detection unit that detects the third point, which is the center point of the mouth, and the detected first to third points are connected to each other to form a triangle. And a control unit for controlling the display unit to determine whether the triangle meets a symmetry criterion and output a determination result.

On the other hand, the control unit first determines whether one side of the triangle connecting the first point and the second point is parallel to the ground, and when it is determined that the one side is parallel to the ground, the left and right on the basis of the center line of the triangle It is possible to determine whether is symmetric.

In this case, the controller may control the display unit to output a message indicating that the calibration of the front image is completed when the left and right are symmetrical with respect to the center of the triangle as a result of the second determination.

Alternatively, the controller may determine that the one side is not parallel to the ground as a result of the first determination, or if the left and right are not symmetrical based on the center of the triangle as a result of the second determination, the correction of the front image is performed. The display unit may be controlled to output a message indicating that it is necessary.

In this case, the controller outputs a message indicating that the front image needs to be corrected, and then reconstructs a triangle from the front image retaken by the camera and determines whether the reshaped triangle meets a symmetry criterion. can do.

The face image photographing apparatus includes eyebrow position information, nose position information, face line information, and ear position information in any one of a first region of a left region and a right region with respect to the center line of the triangle of the front image. The apparatus may further include a second detector configured to detect facial feature points.

In addition, the controller may set one region centered on a point symmetrical to the facial feature point of the first region in the second region, and detect a face matching point that matches within the set region.

In this case, when the face matching point matching the face feature point of the first area is not detected in the second area, the controller calculates a face matching target point in the second area by using the following equation,

Here, c ₁ is a main face feature point in the first area, c ₂ is a first sub face feature point in the first area, c ₃ is a second sub face feature point in the first area, and d ₁ is the _first face feature point in the first area. A facial matching target point to be calculated in area 2, d ₂ is a first sub face matching point matching c ₂ in the second area, and d ₃ is a second sub face matching c ₃ in the second area. Matching point.

On the other hand, the facial feature point detection method of the facial imaging apparatus according to an embodiment of the present invention, the first step of capturing and displaying the front image of the target person in real time using a camera, by the camera using the color information Detecting the first and second points, which are the center points of both eyes, in the photographed front image, and forming a triangle by connecting the second and third points detected to the third point, which is the center point of the mouth. And determining whether the triangle meets a symmetry criterion, and outputting the determination result on the display screen.

The third step is a first process of first determining whether one side of the triangle connecting the first point and the second point is parallel to the ground, and the center line of the triangle when it is determined that the one side is parallel to the ground. As a reference, it may include a second process of determining whether the right and left are symmetric.

In this case, in the third step, as a result of the second determination, if the left and right are symmetric with respect to the center of the triangle, a message indicating that the calibration of the front image is completed may be output on the display screen.

Alternatively, in the third step, when it is determined that the one side is not parallel to the ground as a result of the first determination or as a result of the second determination, it is determined that the left and right are not symmetrical with respect to the center of the triangle. A message indicating that calibration is necessary may be output on the display screen.

In this case, after outputting a message indicating that the front image needs to be corrected, the third step may reconstruct a triangle from the front image re-photographed by the camera, and the reshaped triangle does not meet the symmetry criterion. It can be determined.

The facial feature point detection method includes eyebrow position information, nose position information, face line information, and ear position information in any one of a first region of a left region and a right region with respect to the center line of the triangle in the front image. The method may further include detecting a facial feature point.

The method may further include a fifth step of setting a region around a point symmetrical to the facial feature point of the first region in the second region and detecting a face matching point that matches within the set region. .

In the fifth step, when a face matching point matching the facial feature point of the first area is not detected in the second area, a face matching target point is calculated in the second area by using the following equation.

Here, c ₁ is a main face feature point in the first area, c ₂ is a first sub face feature point in the first area, c ₃ is a second sub face feature point in the first area, and d ₁ is the _first face feature point in the first area. A facial matching target point to be calculated in area 2, d ₂ is a first sub face matching point matching c ₂ in the second area, and d ₃ is a second sub face matching c ₃ in the second area. Matching point.

According to the present invention, it is possible to correct the posture by determining whether the triangle connecting the pupil center point and the mouth center point satisfies the symmetry condition in the front image of the photographed target person. In addition, by detecting a facial feature point using one of the left region and the right region in the front image of the corrected posture, and detecting the face matching point in the symmetrical region, the facial feature point can be detected more quickly and accurately. .

In addition, even if a face matching point is not detected in a symmetrical area, a face matching target point may be calculated to provide a front image in which all facial feature points are displayed, and a doctor may deliver a higher quality diagnosis to a patient during medical diagnosis. do.

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

1 is a block diagram illustrating a facial imaging apparatus according to an exemplary embodiment of the present invention. Referring to FIG. 1, the facial imaging apparatus 100 may include a camera 110, a display 120, a first detector 130, a second detector 140, a controller 150, and a memory 160. It includes more.

The camera 110 captures a front image of the target person and converts the image into an electrical signal. In this case, the type of the camera 110 used is not limited, but generally a digital camera may be used.

The display unit 120 outputs the front image captured by the camera 110.

The first detector 130 detects the first and second points, which are the center points of both eyes, and the third point, which is the center point of the mouth, using color information in the front image captured by the camera 110.

The controller 150 controls the overall operation of the facial image capturing apparatus 100, and in particular, controls the operations of correcting the front image and detecting facial feature points. In this case, the controller 150 controls an operation of calibrating the front image by using the first to third points detected by the first detector 130, and thereafter, the first region detected by the second detector 130. The operation of detecting the facial feature point in the second area is controlled by using the facial feature point of.

First, the front image correcting operation of the controller 150 will be described. The controller 150 forms a triangle by connecting the first to third points detected by the first detector 130. In this case, the first point is the right eye center point (or the left eye center point), the second point is the left eye center point (or the right eye center point), and the third point is the lip center point. Therefore, the side connecting the first point and the third point and the side connecting the second point and the third point have the same length, thereby forming an isosceles triangle.

The controller 150 determines whether the triangle meets the symmetry criteria. This determination is to check whether the face is rotated in the left or right direction in the front image. Specifically, the controller 150 primarily determines whether one side of the triangle connecting the first point and the second point is parallel to the ground. When it is determined that one side is parallel to the ground, the second image determines whether the left and right sides of the front image are symmetrical based on the center line of the triangle. In this case, the center line of the triangle may be a line perpendicular to the ground passing through the third point in the triangle.

The controller 150 controls the display 120 to store the front image in the memory 160 and output a message indicating that the calibration of the front image is completed when the left and right are symmetrical with respect to the center of the triangle.

On the other hand, if one side of the triangle is not parallel to the ground or the left and right are not symmetrical with respect to the center line of the triangle, the controller 150 controls the display unit 120 to output a message indicating that the front image needs correction. . Accordingly, the target person (person) checks the message and corrects the posture so that the front image can be retaken. In this case, the message output through the display unit 120 does not necessarily have to be a text form, but may be an indicator having a specific color.

Subsequently, the controller 150 may reconstruct the triangle using the front image re-photographed by the camera 110 and determine whether the triangle meets the symmetry criteria to obtain the front image of the correct posture.

When the front image of the target person is calibrated through the above-described operation, the controller 150 controls the operation of detecting the facial feature points. Before describing a specific operation thereof, the second detector 140 will be described.

The second detector 140 detects facial feature points including eyebrow position information, nose position information, face line information, and ear position information in the first area. Here, the first region refers to any one of the left region and the right region with respect to the center line of the triangle in the front image of the person. In this case, any one area selected as the first area may be preset, and may be arbitrarily changed every time the facial feature point is detected.

The controller 150 detects face matching points matching the face feature points in the second area by using the face feature points detected in the first area through the second detector 140. The second region may be a region symmetrical with the first region with respect to the center line of the triangle, and the second region may be a right region when the first region is a left region. Therefore, the controller 150 matches the facial feature points detected in the first area to the second area based on the center line of the triangle and detects facial matching points matching each facial feature point. In this case, facial matching points detected in the second area may be facial feature points in the second area.

If the position of the eyebrow is found in the second area by using the position information of the eyebrow among the facial feature points of the first area, the first area and the second area are matched to determine whether there is a point matching the eyebrow position of the first area. do. Specifically, the face matching point for the position of the eyebrows located in the second area is searched on the second area for a point symmetrical to the position of the eyebrows of the first area with respect to the centerline of the triangle, and a rectangle of a predetermined size around the detected point. The area can be formed and detected therein. As a result, the face matching point with respect to the position of the eyebrow in the second area can be detected. As described above, the facial feature points are detected only in the first area, and the facial matching points are detected in the second area by using the facial feature points, thereby making it easier and faster to detect the facial feature points in the front image.

The controller 150 calculates a face matching target point when a face matching point matching the face feature point of the first area is not detected in the second area. In this case, the face matching target point may be calculated using the following equation.

Where c ₁ is the main face feature point in the first area, c ₂ is the first sub face feature point in the first area, c ₃ is the second sub face feature point in the first area, and d ₁ is calculated in the second area The face matching target point, d _2, is a first sub face matching point that matches c ₂ in the second region, and d ₃ is a second sub face matching point that matches c ₃ in the second region. In this case, each point, that is, c ₁ , c ₂ , c ₃ , d ₁ , d _2, and d ₃ may be represented by two-dimensional vectors as two-dimensional points.

On the other hand, by forming a triangle in the front image to determine whether the symmetry criteria are met, the left region and the right region based on the center line of the face in the front image does not achieve perfect symmetry. Therefore, since there is a slight asymmetry, the face matching target point can be calculated by compensating for the asymmetry between the left region and the right region through a proportional constant in consideration of this.

When the face matching target point is calculated by the above equation, the controller 150 interpolates the area of the face matching target point to correspond to the area of the face feature point that is matched in the first area.

2 is a schematic diagram illustrating a front image calibration process according to an embodiment of the present invention. Referring to FIG. 2, a first point P ₁ , a second point P ₂ , and a third point P ₃ are detected in the front image 200 captured by the camera 110. In this case, the first and second points P ₁ and P ₂ are centers of both pupils, and the third point P ₃ is a center point of the mouth. Detection of the first to third points P ₁ , P ₂ , and P ₃ may be performed using color information in the front image 200.

In detail, the face area is detected in the front image 200. In general, a face region is detected by searching for a skin color in each pixel of the front image by using a point where a lot of skin color (skin color) is distributed in the face region.

Then, both eye regions are detected in the face region. In general, the eye region is black, so that each pixel of the face region may be searched for black to detect the first point P ₁ and the second point P ₂ , which are the center points of both eyes. In this way it is possible to detect the lip region and to detect the center point at the third end point P ₃ at both ends of the lip. In this case, the method of detecting the pupil area and the lip area in the face area may further use general location information of the components constituting the face, in addition to using the color information.

When the first to third points P ₁ , P ₂ , and P ₃ are detected, the first point P ₁ , the second point P ₂ , the second point P ₂ , and the third point P ₃ ) Then, the third point P ₃ and the first point P ₁ are connected to form a triangle. In addition, it is first determined whether one side L connecting the first point P ₁ and the second point P ₂ in the triangle is parallel to the ground (here, the horizontal line H of the front image). In this case, it is preferable that one side L of the triangle and the ground be exactly parallel, but it can be determined to be parallel even when the error range is within ± 1 to 5%.

As a result of the first determination, when it is determined that one side L connecting the first point P ₁ and the second point P ₂ is parallel to the ground H, the center line C1 of the triangle is referred to. The second judgment is whether the left and right triangles are symmetrical. The left and right symmetry of the triangle can be determined by checking whether the positions of both sides of the triangle and the positions of the vertices coincide with each other based on the center line of the triangle. In this case, it is preferable that the positions of both sides of the triangle and the positions of the vertices are exactly the same, but it can also be determined that the case also has an error range within ± 1-5%.

As a result of the second determination, when the left and right sides of the triangle are symmetrical, the front image is corrected, and a message indicating that the front image is corrected is output on the display screen.

On the other hand, if one side L connecting the first point P ₁ and the second point P _{2 in the} triangle is not parallel to the ground or is not symmetrical with respect to the center line of the triangle, the display screen A message indicating that the front image needs to be corrected is output on the screen so that the front image is retaken. The front image is corrected by re-determining whether the triangle meets the symmetry criterion using the rephotographed front image.

3A to 3C are schematic diagrams illustrating a facial feature point detection process according to an embodiment of the present invention. In FIGS. 3A to 3C, the center lines C2 and C3 are the same as the center line C1 of the triangle shown in FIG. 2, but the triangle is not shown for clarity.

3A is a diagram illustrating a process of detecting a face matching point in a second area by using face feature points detected in the first area. Referring to FIG. 3A, the front image 300 calibrated based on the center line C2 is divided into a first area A and a second area B. FIG. In this case, the first area A may be a left area in the front image 300, and the second area B may be a right area.

The facial imaging apparatus 100 detects a facial feature point including ear position information a ₁ and a ₃ , eyebrow position information a ₂ , and nose position information a ₄ in the first area A. FIG. . Then, the face matching points are detected by matching the facial feature points of the first area A with the second area B by symmetrically matching the center line C2. Specifically, as shown in FIG. 3A, the point corresponding to the eyebrow position information a ₂ of the first area A based on the center line C2 of the first area A is defined as the second area B. As shown in FIG. Search on. Then, a rectangular region (dotted display region) having a predetermined size is formed around the searched point, and the face matching point b ₂ can be detected therein. Using this method, face matching points b ₁ and b ₃ matching ear position information a ₁ and a ₃ in the first area A and face matching points matching nose position information a ₄ are matched. (b ₄ ) can be further detected.

3B and 3C illustrate a process of detecting a face matching point when the face matching point is not detected in the second region. Referring to FIG. 3B, an image damage area R exists in the front image 400.

When photographing the front image 400, some regions may not be photographed due to an external environment or image noise. Accordingly, the image damage area R may be included in the front image 400. When the image damage area R is included in the front image 400, the facial feature may not be clearly output. Therefore, a process of detecting the facial feature points of the image damage area R is necessary.

First, facial feature points c ₁ , c ₂ , c ₃ , and c ₄ are detected in the first region C. FIG. Then, the face matching points are detected by matching the facial feature points of the first area C with the second area D based on the center line C3. In this case, the face matching points d ₂ , d _3, and d ₄ matching the facial feature points c ₂ , c _3, and c ₄ of the first region C were detected, but matched to c ₁ by the image damage region R. No facial matching point was detected. Therefore, setting the face matching points that are not detected by the face matching the target point (d ₁₎ and by using Equation 1 calculates the face matching the target point (d _1).

In order to calculate the face matching target point d ₁ , necessary points are set. Specifically, since the facial feature point c ₁ of the first region C matches the facial matching target point d ₁ to be calculated, the facial feature point c ₁ is set as the main facial feature point, and c ₂ and Set c ₃ as the first and second sub facial feature points. Further, d ₂ and d ₃ , which match each of the first and second sub face feature points, are set as the first and second sub face matching points, respectively. FIG. 3C illustrates only points necessary for calculating the face matching target point d ₁ by deleting the front image in FIG. 3B.

The face matching target point d ₁ is calculated by substituting the coordinate values of the respective points set as described above in Equation 1. In this case, the x-coordinate value in the coordinate value of each point has a positive increase value on the right side and the negative decrease value on the left side of the center line C3. The y coordinate values all have the same directionality. In addition, when the coordinate value of each point is substituted in Equation 1, the x coordinate value may be substituted for the absolute value.

In this manner, the face matching target point d ₁ located in the image damage area R is calculated and interpolated. Specifically, the front image (interface image) is interpolated by interpolating pixel values of the face matching target point d1 to correspond to the pixel value of the facial feature point c ₁ of the first area C matching the face matching target point d ₁ . 400, all facial feature points can be displayed.

FIG. 4 is a flowchart illustrating a method for detecting facial feature points of the facial imaging apparatus illustrated in FIG. 1. First, the facial image capturing apparatus 100 captures a front image of a target person (person) (S510).

Next, the facial image capturing apparatus 100 detects first to third points in the captured front image (S515). Specifically, the first and second points, which are the center points of both eyes, are detected in the front image, and the third point, which is the center point of the mouth, is detected. Then, the first to third points are connected to each other to form a triangle (S520).

Thereafter, the facial imaging apparatus 100 determines whether triangles formed by connecting the first to third points with each other meet the symmetry standard (S525). Specifically, it is determined whether one side of the triangle, that is, the side connecting the first and second points is parallel to the ground, and in the case of the parallel, it is determined whether the left and right are symmetric based on the triangle center line.

As a result of the determination in S525, when the triangle does not meet the symmetry criterion, a message requesting correction of the front image is output on the display screen (S535). Then, the corrected front image is retaken (S540), and steps S515 to S525 are performed again.

If it is determined by the above-described process that the triangle formed in the front image meets the symmetry standard, a message indicating that the calibration for the front image is completed is output on the display screen (S530).

Meanwhile, the facial imaging apparatus 100 detects facial feature points using the calibrated front image. In detail, the facial feature point in the first area is detected based on the center line of the triangle in the front image (S545). In this case, the first region may be one of a left region and a right region based on the center point of the triangle.

Next, in a second area symmetrical to the first area, a face matching point matching the face feature point of the second area is detected (S550). If all face matching points matching the face feature points of the first region are detected through this process (S555), the face feature point detection is terminated.

However, when all of the face matching points are not detected in the second area (S555), a face matching target point matching the face feature point of the first area is detected (S560). In this case, the face matching point may be detected by Equation 1 described above. When the facial matching target point is detected, the interpolation is performed to correspond to the corresponding facial feature point. By using the above method, facial feature points can be detected more quickly.

Although the above has been illustrated and described with respect to the preferred embodiments of the present invention, the present invention is not limited to the above-described specific embodiments, it is common in the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

1 is a block diagram showing a face image photographing apparatus according to an embodiment of the present invention;

2 is a schematic diagram showing a front image correcting process according to an embodiment of the present invention;

3A to 3C are schematic diagrams illustrating a facial feature point detection process according to an embodiment of the present invention;

FIG. 4 is a flowchart illustrating a method for detecting facial feature points of the facial imaging apparatus illustrated in FIG. 1.

<Description of Signs for Main Parts of Drawings>

100: facial video recording device 110: camera

120: display unit 130: first detection unit

140: second detection unit 150: control unit

160: memory

Claims (16)

A camera for photographing the front image of the target person in real time; A display unit for outputting a front image captured by the camera as a color image; After detecting the face region of the target person in the front image by using the color information of the face, detecting both eyes and the mouth in the detected face region, the first and second points which are the center points of both eyes and the mouth A first detector detecting a third point, the center point; And, The detected first to third points are connected to each other to form a triangle, and a first determination is made that one side of the triangle connecting the first point and the second point is parallel to the ground, and the one side is parallel to the ground. If it is determined that the triangle is passed to the left and right based on the center line of the triangle which is a line perpendicular to the ground passing through the third point of the triangle to determine whether the left and right of the triangle is symmetrical, and outputs the determination result Face imaging apparatus including a control unit for controlling the display unit. delete The method of claim 1, The control unit, And as a result of the second determination, if the left and right sides of the triangle are symmetrical, the display unit controls the display unit to output a message indicating that the calibration of the front image is completed. The method of claim 1, The control unit, The display is configured to output a message indicating that the front image needs to be calibrated when it is determined that the one side is not parallel to the ground as a result of the first determination or that the left and right of the triangle are not symmetrical as a result of the second determination. Facial imaging apparatus, characterized in that for controlling the wealth. The method of claim 4, wherein The control unit, After outputting a message indicating that the front image needs correction, reforming a triangle from the front image re-photographed by the camera and performing the first and second judgment using the reshaped triangle. Facial imaging device characterized in that. The method of claim 1, When the face region is divided into a left region and a right region by the center line of the triangle, one of the left region and the right region is set as a first region, and the eyebrow position information, nose position information, and face in the set first region. And a second detector for detecting facial feature points including line information and ear position information. The method of claim 6, The control unit, By setting any one of the left area and the right area not set as the first area as a second area, a predetermined area is set based on a point symmetrical to the facial feature point of the first area in the set second area, And a facial matching point detected within a predetermined predetermined area. The method of claim 7, wherein The control unit, In the second area, when a face matching point matching the facial feature point of the first area is not detected, a face matching target point is calculated in the second area by using any one of Equations 1 and 2 below.

… Equation 1

… Equation 2 Here, c ₁ is a main face feature point in the first area, c ₂ is a first sub face feature point in the first area, c ₃ is a second sub face feature point in the first area, and d ₁ is the _first face feature point in the first area. A facial matching target point to be calculated in area 2, d ₂ is a first sub face matching point matching c ₂ in the second area, and d ₃ is a second sub face matching c ₃ in the second area. Facial imaging apparatus characterized in that the matching point. A first step of capturing a front image of a target person in real time using a camera and outputting the front image as a color image on a display screen; The face region of the target person is detected from the photographed front image by using color information of the face, both eyes and the mouth are detected from the detected face region, and then the first and second points which are the center points of both eyes and the Detecting a third point, which is a center point of the mouth; Forming a triangle by connecting the detected first to third points, and primarily determining whether one side of the triangle connecting the first point and the second point is parallel to the ground, the one side is parallel to the ground When it is determined, the triangle is divided into left and right directions based on the center line of the triangle, which is a line perpendicular to the ground passing through the third point of the triangle, and secondly determines whether the left and right sides of the triangle are symmetric, and displays the determination result. A third step of outputting through a screen; And If the left and right sides of the triangle are symmetrical as a result of the second determination, one of the left region and the right region is set as the first region in the face region divided into the left region and the right region by the center line of the triangle, and the set first And a fourth step of detecting a facial feature point including eyebrow position information, nose position information, face line information, and ear position information in one region. delete The method of claim 9, The third step, And if the left and right sides of the triangle are symmetrical as a result of the second determination, a message indicating that the calibration of the front image is completed is output to the display screen. The method of claim 9, The third step, If it is determined that one side is not parallel to the ground as a result of the first determination, or that the left and right of the triangle are not symmetrical as a result of the second determination, a message indicating that the front image needs to be corrected is displayed on the display screen. A facial feature point detection method of a facial imaging apparatus, characterized in that output. The method of claim 12, The third step, After outputting a message indicating that the front image needs correction, reforming a triangle from the front image re-photographed by the camera and performing the first and second judgment using the reshaped triangle. A facial feature point detection method of a facial imaging apparatus. delete The method of claim 9, By setting any one of the left area and the right area not set as the first area as a second area, a predetermined area is set around the point that is symmetrical to the facial feature of the first area in the set second area. And a fifth step of detecting a face matching point matched within a predetermined predetermined area. The method of claim 15, The fifth step, In the second area, when a face matching point matching the facial feature point of the first area is not detected, a face matching target point is calculated in the second area by using any one of Equations 1 and 2 below.

… Equation 1

… Equation 2 Here, c ₁ is a main face feature point in the first area, c ₂ is a first sub face feature point in the first area, c ₃ is a second sub face feature point in the first area, and d ₁ is the _first face feature point in the first area. A facial matching target point to be calculated in area 2, d ₂ is a first sub face matching point matching c ₂ in the second area, and d ₃ is a second sub face matching c ₃ in the second area. A facial feature point detection method of a facial imaging apparatus, characterized in that the matching point.

Images